Of What Use, Emotion?
I found myself this week with a fast-approaching deadline and no real purpose or direction to my writing when I read this article by the inestimable Matt Waldman. In it, Waldman talks about the role that emotion plays in football and the ways that analysts on different sides of the issue talk past each other.
Now, I doubt I really need to spend much time singing the praises of Matt Waldman. If you know who I am, you doubtless know who Matt Waldman is too, (though the same could not be said of the inverse). If by some chance you happen to be unfamiliar with him, I would strongly advise against reading anything he ever writes unless you happen to like thorough, thoughtful, cogent, nuanced, and well-articulated analysis.
Anyway, while reading his article, I found myself agreeing with every point he raised. At the same time, I found myself agreeing with those to whom his article was responding. How could this possibly be?
The Very Real Impact of the Contract Year
A common belief among football fans is the motivating power of a “contract year”. The idea is that when they are in the final year of their current contract, NFL players give a little bit more effort knowing that their next big payday depends on their current production.
Is the “contract year effect” a real thing? Of course it is, and I challenge anyone to explain the career of Albert Haynesworth without it. For those who don’t remember, through the first five years of his career Haynesworth was a solid defensive tackle who distinguished himself more for an incident where he stomped on the helmetless head of an opposing lineman than for his actual play. In 2007, in the last year of his contract, Haynesworth unexpectedly became one of the best defensive linemen in the league, doubling his previous career high in sacks and anchoring one of the better defenses in the NFL. After the season, he was named a first-team AP All Pro.
After the season was over, Tennessee used the franchise tag to retain Haynesworth’s rights for another year, and Haynesworth somehow managed to raise his level of play even higher. Haynesworth was the most dominant presence in the league and his excellence propelled the Titans to the best record in the NFL.
After two straight “contract years”, Haynesworth left Tennessee in free agency and signed a blockbuster contract with Washington that, at the time, made him the highest-paid defensive player in history.
Immediately Haynesworth’s level of play fell off a cliff. He showed up to camps overweight and took several weeks before he was able to pass the team’s basic conditioning test. Those early struggles were a harbinger of things to come, and the team eventually suspended him before trading him. Haynesworth played just 20 games for Washington.
Everyone who watched the Haynesworth saga unfold could be absolutely certain about one thing: in Tennessee, Haynesworth was motivated; in Washington, he was not.
The Completely Fabricated Impact of the Contract Year
So obviously the “contract year effect” is a real thing and players play harder when they know their next big payday is on the line. Except the thing is… they don’t.
If you look at the big picture, there’s little to no difference between player performance before and after signing a big contract. Typically players who sign a big contract see a decline in performance, but that’s because players who earn a big contract typically are coming off of great seasons, and players who are coming off of great seasons typically regress to the mean.
If you compare the cohort of players who just signed a new contract with a cohort who performed similarly but did not receive a new contract, the decline is virtually identical. And this makes sense. The very idea of “contract years” misunderstands how franchises operate.
If a player is performing poorly, teams can and frequently will part ways with them. In this sense, every single season is potentially the final season of a player’s contract. The idea that we can predict which players will overperform and which will underperform based on the timing and size of their last contract simply doesn’t hold water.
So which is it? Is the contract year effect a very real phenomenon, like everyone who saw Albert Haynesworth play knows it to be? Or is it a false narrative, like everyone who looks at the data knows it to be?
The truth is that it’s both. Beliefs are capable of being either descriptive or predictive. As a descriptive belief, the idea that some players raise their play when a contract is on the line is easily shown to be true. As a predictive belief, the idea that we should expect players to raise their play when a contract is on the line is easily demonstrated to be false.
Grief, Fear, And Other Matters of Great Import
This is hardly the only example of such a contradiction. On December 21st, 2003, Brett Favre’s father died tragically at 58. On December 22nd, 2003, Brett Favre exorcised his personal demons on national television against the Oakland Raiders, throwing for four first-half touchdowns and finishing with 399 yards on just 30 attempts.
The game remains the highest-rated game in Favre’s career by passer rating, though that’s not what everyone remembers. Instead, the raw cathartic emotion Favre exhibited on the field and on the sidelines became seared into the national consciousness.
Favre’s battles against his own personal demons were so apparent that even the Raiders faithful in attendance were brought to their feet.
Ten years later, on October 11th 2013, Adrian Peterson’s 2-year-old son was tragically beaten to death by his mother’s boyfriend. Two days later, Peterson suited up and took the field against the Carolina Panthers. Peterson gained 83 yards from scrimmage, with over half of that total coming on two plays after the Panthers had already raced to a four-score lead.
Brett Favre’s performance was undeniably affected by the swirl of emotions he doubtless felt that night. Adrian Peterson’s was, too. As was the performance of every other athlete in the history of the NFL who has returned to the field while still sorting through the fallout of intensely personal loss.
One could look at these instances and others like them and conclude that raw emotion has a substantial effect on player performance. One could look at these instances and others like them and conclude that raw emotion has no measurable effect on player performance. Both conclusions are valid and supported by the evidence. One belief invokes a description of events that have already happened. The other creates a prediction of events yet to come.
A tragic loss could bring out the best in a player, or it could bring out the worst. Triumph can lead to success or failure. Fear and doubt can inspire greatness or elicit collapse. Sometimes they can do none of these things. Other times, they can do all of them. And it's not just different players who react differently; even the same player will react differently from situation to situation.
A player who collapses against an opponent might find a new wellspring of resolve for the rematch. Or he might relive his failures until they consume him and failure becomes a self-fulfilling prophecy. Or maybe the collapse was never a collapse in the first place, it was an illusion created by our need to write meaningful narratives out of meaningless sample sizes.
The impact of emotion can be seen, but only with the benefit of hindsight.
Descriptive vs. Predictive Beliefs
This dichotomy of descriptive vs. predictive beliefs is at the heart of the argument between the two camps. One side tells a story of what happened. The other tells a story of events yet to come. Together they provide non-overlapping magisteria through which we can understand football.
Descriptive beliefs are welcomed houseguests and we are their appreciative friend. Their presence enriches our lives, and their worth is demonstrated by their ability to warm and fill a cold and empty mental landscape.
It’s possible for these beliefs to overstay their welcome, to linger long enough that their endearing ticks become annoying and their shoddy hygiene begins to outweigh their dazzling companionship. At such a time, we kindly ask these beliefs to move on, but we retain our fond memories of their stay.
Predictive beliefs are tenants and we are their responsible landlord. Their presence enriches our bank account and frees up time for us to spend as we see fit. Our personal feelings towards them are of little matter so long as their rent is not overdue.
Predictive beliefs pay rent by anticipating outcomes, and as long as the rent is on time every month, we are better off for their presence. When they become consistently delinquent, however, it is incumbent upon us to evict them and replace them with new beliefs capable of keeping up with their responsibilities. At such a time, we feel no regret or sense of nostalgia, but instead a hope that the new tenant is a good one and makes our life easier.
A man with dozens of tenants but no welcomed houseguests has plenty of riches, but no one to spend them with. A man with dozens of houseguests but no tenants has all the company he could want, but no way to keep the utilities on. Instead, a full life requires a healthy mix of both.
I believe that John Elway was a better quarterback than Joe Montana. This is a descriptive belief, one that makes no attempt to predict any results, one that doubtless says as much about myself as it does about Elway or Montana.
I believe that Peyton Manning is devastatingly effective in close games. This is a predictive belief, one that doesn’t just describe past results but assigns probabilities to the future.
When proper distinctions are not made between the two types of belief, confusion can result. You would no more discuss the last time your houseguest paid his rent than you could reminisce about your tenant’s last birthday.
Similarly, discussion of whether Tony Romo choked in the playoffs as a first-year starter is categorically different than discussion of whether Tony Romo is a choker.
The first discussion involves a story about events of the past, a uniquely human conceit. From the earliest moments of our species’ history we have been a tribe of storytellers, scrawling our experiences on the walls of caves, coding our mythologies in the stars, telling our tales on yellowed pages and silver screens.
Indeed, while storytelling is central to every facet of human existence, it is somehow more central to football than perhaps any other. While basketball has Oscar’s triple-doubles and baseball offers Cal’s 2632, Football gives us the autumn wind and the frozen tundra. Thanks to the tireless work of Steve Sabol, Ed Sabol, and John Facenda, previous generations of football have bequeathed us a history that is far more narrative than numerical.
It is into that context that we must place Romo's performance, finding a place for his story among the many threads that have been woven into the NFL's tapestry through the years. Dan Marino pretending to spike a football. Earl Campbell running through a defender. A young Randy Moss seeking vengeance against the Cowboys. Buddy Ryan punching Kevin Gilbride in the jaw. Tony Romo botching the hold against the Seahawks.
The question of whether Tony Romo is a choker, however, is altogether different. Choking is a one-time event, whereas a choker is someone for whom choking is destined to happen again and again. The term choker carries with it a set of expectations and anticipations.
In short, the belief that Tony Romo is a choker is one that should pay rent and, to put it bluntly, the check is long overdue. This week’s game against the Seahawks was merely the latest in a long list of examples where the implicit prediction failed to pay.
Is choking real? I would invite anyone who doubts it to google “Jana Novotna” some time. Anyone who has ever set out in any capacity with the weight of expectations on his or her shoulders understands the looming specter of collapse that comes with that burden.
Are chokers real? I would venture that there’s precious little evidence from the NFL suggesting as much. Even Jana Novotna eventually won at Wimbledon.
Unfortunately, this leaves us in a rather uncomfortable place. We must acknowledge that momentum is a very real force that can only be appreciated after it’s gone. The 2012 Denver Broncos came back from a 24-point deficit against the Chargers and wound up winning by 11; each success built on itself as Denver’s momentum became self-sustaining. The 2012 New England Patriots erased a 21-point deficit against the 49ers and wound up losing by 7; any momentum they might have had was, with the benefit of hindsight, purely illusory as an emotionally-resilient 49ers squad weathered the storm and regained control.
Emotion will play a large role in determining the outcomes of football games, but we can’t know what that role will be until we are able to perform an autopsy, long after such knowledge would have done us any good.
Is Kobe Bryant clutch? That depends on whether we’re asking about the shots he’s made or the one he’s about to take. Is Antone Smith an unstoppable big-play machine? It's a question of whether we're looking at the games he's already played or the ones remaining on his schedule. Is Rob Gronkowski injury prone? Well, are we discussing the time he has missed in the past or the unknown amount of time he will miss in the future?
It’s important to critically examine our beliefs and determine which ones make claims about the future. If a belief purports to give us insight into how events will unfold, we must ensure it is making good on its rent, or else we must evict it. If a belief makes no claims about the future, though, then it would be crass to ask it to pay its way. Instead, it is to be treasured like an old friend and held close for as long as it leaves us better off than we were without it. After all, what is fandom if not merely a series of descriptive beliefs— beliefs that our team is somehow better, more meritorious, more worthy of our adoration.
A belief like that is always welcome, and our lives are all the richer for it. Just because it’s wrong doesn’t mean it’s not true.